Beyond Words: Linguists Can Change the Game in E-Discovery

A look at the importance and nuances of linguistics in in e-discovery, litigation, and, searching through ever growing tranches of ESI.

, Legaltech News

   | 3 Comments

A look at the importance and nuances of linguistics in in e-discovery, litigation, and, searching through ever growing tranches of ESI.

This premium content is reserved for Legal Technology subscribers.

Continue reading by getting started with a subscription.

Already a subscriber? Log in now

What's being said

  • Michael Morneault

    Excellent comments. And we especially agree with your assertion, David; what TREC and other adjudications have shown consistently is that the right experts (including, but not limited to, linguists) leveraging complementary state-of-art technologies to effect proven processes can achieve outstanding client results. It‘s this combination of contributions that leads to particularly potent results.

  • David Kittrell

    A good discussion. The techniques described however are not limited to linguists. Linguists and non-linguists have been used by various firms to improve criteria lists. Frankly, the big problem I‘ve seen is convincing counsel of the need to do such processes, not whether linguists or simply a structured criteria development process in conjunction with typical end user research. Another issue is a possible consequence of adding or improving criteria: more document hits. Counsel often expect better criteria to reduce the result set size while often the opposite is true. There are ways to mitigate this but it‘s critical that counsel understand -- and preferably, expect -- such outcomes. Again, a structured approach to criteria development can mitigate precision/recall issues and explain the results but it can be non-intuitive for naive counsel. Machine learning, statistical sampling and iterative validation can help but there‘s no magic perfect approach to every scenario. Bottom line: the process is more important than staff -- linguists are useful but similar, sometimes even better, outcomes can be achieved with research analysts employing the right process and tools. I see articles such as this being a good reminder of what Blair and Maron pointed out 30 years ago: ad hoc criteria development seldom produces desired recall/precision results. Tools and processes employed in non-litigation domains can greatly improve on counsel‘s traditional methodology. But getting counsel to look at alternatives -- whether the newest shiny TAR n.x appoach or simply a common-sense, structured, approach to a mix of tools and techniques -- has taken over twenty years despite project evidence there are better ways.

  • Bill Speros

    Thank you, Ms. Podolny, for so concisely and capably describing a topic that is critical but oftentimes (and at great peril) ignored by litigators.It brought to mind this: A decade ago we had to find from among a million electronic documents all that referred to people who had “not reached the age of majority” and, therefore, were “minors.” What we found is this: the attorneys’ perspective was blind to what words, terms of art, and jargon the document authors’ employed. More specifically, nearly all references to the attorneys’ key keyword, “minor,” pertained not to age but to the severity of injury. Instead, to identify relevant documents we quickly interviewed the authors and based upon their natural language, their business problem and tendencies, we searched (reiteratively) for key terms—using all the necessary “nyms”—that referred to, for example, young peoples’:•activities (e.g., school, [various sports], bicycle, crossing guard)•medical treatment (e.g., pediatrician, [local children’s hospitals names])•relationships (e.g., mother, father, aunt, uncle, brother, sister)…That social-linguistic bird-dog pointed to documents that carried the scent of young people.Consequently, we found what your thoughtful article asserts: that whereas litigators are subject matter experts regarding legal claims and defenses, it is not litigators but the documents’ authors who are subject matter experts regarding how within the prospectively responsive documents relevant ideas are expressed.

Comments are not moderated. To report offensive comments, click here.

Preparing comment abuse report for Article #1202756531694

Thank you!

This article's comments will be reviewed.